MOLECULAR BIOLOGY AND PHYSIOLOGY. Seed-Fiber Ratio, Seed Index, and Seed Tissue and Compositional Properties of Current Cotton Cultivars

Transcription

1 The Journal of Cotton Science 22:60 74 (2018) The Cotton Foundation MOLECULAR BIOLOGY AND PHYSIOLOGY Seed-Fiber Ratio, Seed Index, and Seed Tissue and Compositional Properties of Current Cotton Cultivars Michael K. Dowd*, Scott M. Pelitire, and Christopher D. Delhom ABSTRACT Because of the continual efforts to breed cotton for increased fiber yield, several seed/fiber compositional properties have likely shifted over the decades. Conversations with breeders, ginners, and oil processers have identified several concerns, including smaller seed size, weaker hulls, increased seed and hull fragment contamination of fiber, and reduced seed oil and protein levels all of which directly affect the economic value of the crop. To better understand these changes, field cotton samples of current cultivars were collected from areas around Stoneville, MS; Lubbock, TX; and Las Cruces, NM. The samples were ginned and cleaned to determine seed-to-fiber ratio, seed index, and the proportions of linter, hull, and kernel tissues. Kernels were then analyzed for oil, protein, and gossypol. Results from the threeyear study (2014 through 2016) indicated that the average seed-to-fiber ratio was 1.41 ± 0.11 (range: , as is basis) and has declined compared with data sets published prior to Of the varieties included in the study, seed index averaged 9.75 ± 0.99 g (range: g, as is basis) and also showed an overall decline compared with early published data. Seed tissue proportions have changed less, although a decrease in the percentage of linters was apparent. The average level of seed oil and protein does not appear to have changed much over the years, although oil levels were very low for a few individual cultivars. Over many decades, selective breeding of the cotton plant has resulted in dramatically increased fiber yields. These efforts, however, may have come with consequences for the seed, and a number of complaints have been voiced by ginners M.K. Dowd*, S.M. Pelitire, and C.D. Delhom, Southern Regional Research Center, 1100 Robert E. Lee Blvd., ARS, USDA, New Orleans, LA *Corresponding author: michael.dowd@ars.usda.gov and oil processors regarding changes in seed quality. Weaker seed, smaller seed size, and reduced oil and protein levels have all been discussed. Seed quality affects ginners, textile manufacturers, and oil processers in a number of ways. Weak seed is a problem for both fiber and seed processors, as ginning and handling operations result in hull fragments that contaminate the fiber making it more difficult to clean and process. Concomitantly, damaged seed is more prone to moisture uptake and oil degradation during storage, which causes a direct loss of extractable oil and oil refining problems. Small seed size is a special concern for ginners, as the seed is usually taken as the payment for ginning. In addition to less seed produced per bale of ginned fiber, losses might be amplified if whole seed and seed pieces exit the gin stand with the motes or in the fiber cleaning operations. Small seed size also affects the oil processor s ability to recover the linters (small unginnable fibers) and dehull the seed. Low oil levels are a direct loss for the oil processors, and low protein levels reduce the amount of hulls that can be left with the kernels during oil extraction, making it more costly to prepare a standard 41% protein meal. Values for seed properties were frequently obtained in the early years of studying oil processing, and a number of these studies are discussed in chapters in A.E. Bailey s book Cottonseed and Cottonseed Products (see e.g., chapters by Lund, 1948; and Tharp, 1948). However, since this publication, limited studies have focused on these properties, although the expectations are that some of these values have changed due to the intense breeding for increased fiber yield. To try to better understand the magnitude of these differences, this survey was undertaken to derive seed property values for current commercial cultivars and to compare these values with those discussed in the early literature. METHODS Field cotton samples were collected with the help of the Agricultural Research Service (ARS) cotton gin laboratories located in Stoneville, MS, Lubbock, TX, and Las Cruces, NM. After removing the bulk of

2 DOWD ET AL.: SEED PROPERTIES AND COMPOSITIONAL ANALYSIS OF CURRENT COTTON CULTIVARS 61 the stem, burr, and leaf debris, three pounds of each sample was ginned on a ten-saw laboratory-scale gin, and the fiber and seed fractions were collected and weighed. Fiber moisture was determined by a modified ARS method for ginned fiber (ARS, 1972). Five grams of each fiber sample was weighed into a tall weighing bottle and dried in a convection oven at 105 C for 90 min. The bottles were then capped and placed in a room temperature desiccator (20 min) before weighing. Seed moisture was determined by the American Oil Chemists Society (AOCS) Official Method Aa 3-38 (AOCS, 1998) on ten g samples, which were dried at 130 C for three hr, then allowed to reach room temperature in a desiccator (20 min) before weighing. The fiber was cleaned by two passes on a Shirley Analyzer (Manchester, UK). Seed-to-fiber ratio was calculated after accounting for the trash level determined by the cleaner, and seed index (gram weight of 100 seed) was determined by weighing. Seed were dissected to determine the percentage of linters (i.e., short unginnable fibers), hulls, and kernels. Whole seed was analyzed for linter content by AOCS Official Method Aa 7-55 (AOCS, 1998), which uses a modest acid treatment followed by agitation to brush the linters free of the seed. The procedure determines the linter content by weight difference, and then corrects the moisture to an 8% level (typical of fiber and linter moisture levels at ambient conditions). Approximately 20 g of fuzzy whole seed was then hand cut to separate the kernel fraction from the hull linter fraction, and both fractions were weighed. The percentage of hulls was then determined by difference. Protein and moisture were measured on the hull linter and kernel fractions, and the oil, protein, gossypol, and moisture contents were measured on the kernel tissue. For these tests, the hull linter samples were ground with a Wiley laboratory mill, and the kernels were chopped with a Braun hand-held food chopper, both samples ground to pass through a 20 mesh sieve. Gossypol was measured by high pressure liquid chromatography with the AOCS Recommended Practice Ba 8-99, and protein was determined by combustion with a LECO (St. Joseph, MI, USA) model FP-628 Truspec nitrogen analyzer. Both measurements were made on an as is basis. Moisture was determined on the ground kernel and hull linter fractions by the AOCS Official Method Ba 2a-38 for cottonseed meal (oven drying at 130 C for two hr), which was used to calculate protein and gossypol values of these fractions on a moisture free basis. To determine the crude oil level in kernels, ground kernel tissue was first freeze-dried with a benchtop lyophilizer for three to four days. The oil was then extracted with a Foss (Eden Prairie, MN, USA) Soxtec analyzer with petroleum ether (three g samples, 15 min immersion cycle and three hr extraction cycle), and the recovered oil was determined by weighing. The kernel moisture level was then used to calculate the results on an as is basis. Seed gossypol level was based on the kernel results as there is no detectable gossypol in either the hulls or the fiber. Seed oil was determined from the kernel oil values and the percentage of seed kernel tissue with a small adjustment to account for the expected level of extractable lipid in the linters and hulls, which amounts to less than 1% in each tissue (Tharp, 1948). Seed protein was calculated from the protein levels measured for both the hull linter and kernel tissues. All analyses were conducted on three replicates of each cultivar. Most analytical determinations were made in duplicate, with the average taken as the value for the replicate. Seed index was averaged from three 100 seed weight determinations for each replicate. For each property measured, the data was analyzed for year and location differences with ANOVA and the Tukey mean comparison test (α = 0.05). All results were obtained on an as is basis and on a moisture free (dry weight) basis. Unless otherwise stated, values discussed in the work refer to an as is basis. Individual cultivar values on an as is basis are provided in Supplemental Tables S1- S9. Comparable values on a dry weight basis are provided in Supplemental Tables S10 S18. RESULTS A total of eleven field cotton samples were collected for the 2014 season. One Gossypium barbadense cultivar (DP340) was included. One G. hirsutum cultivar, STV5458, was collected from two locations. All were grown under irrigated conditions. In 2015, thirteen samples were analyzed. Three of these were grown in dryland conditions; the others were grown under irrigation. The STV4946 and DP1044 cultivars were grown in two locations. In 2016, 22 samples were collected from the three regions. All samples were grown under irrigated conditions. One sample, DP1522, was grown in both the Las Cruces and Lubbock regions, and DP340 Pima cultivar was again collected from Las Cruces.

3 JOURNAL OF COTTON SCIENCE, Volume 22, Issue 1, 2018 Production statistics indicated that the cultivars tested contributed significant United States (U.S.) acreage over the three years of study (USDA, 2014, 2015, 2016). The cultivars represented 26.5% of the acres planted in 2014, 27.8% of the acres planted in 2015, and 48.9% of the acres planted in Hence, these samples should be fairly representative of the cotton cultivars seen by ginners, although the set is limited in that only three regions were sampled and the number of samples from each region was variable. A few location and year differences were detected by ANOVA analysis. When evaluated by location, samples from Las Cruces had a higher seed-to-fiber ratio than the other two locations, and samples from Lubbock had a smaller average seed index compared with the other locations (Table 1). When evaluated by year, 2016 samples had a significantly lower seed-tofiber ratio than did samples from the prior two years, and 2014 had a higher average seed index compared with the latter two years. There were no differences in the percentage of linters by year (Table 2), but Las Cruces had increased linters compared to the other two locations. Compared with the earlier years, the percentage of seed hull was greater in 2016, and this occurred with a concomitant decrease in the percentage of kernel tissue. Oil and gossypol levels were significantly reduced for Stoneville produced seed compared with the other locations (Table 3), and correspondingly, the protein levels were higher for this location. Table 1. Seed-to-fiber ratio and seed index summary statistics for cotton varieties produced from (as is basis) z Seed-to-fiber ratio Seed index, g Complete sample population ( ) Ave. ± std. dev ± ± 0.99 Low High ANOVA by year ± 0.15 A 10.4 ± 1.2 A ± 0.10 A 9.48 ± 0.79 B ± 0.09 B 9.58 ± 0.85 B ANOVA by location Las Cruces 1.49 ± 0.14 A 10.3 ± 1.2 A Lubbock 1.38 ± 0.07 B 9.24 ± 0.65 B Stoneville 1.42 ± 0.11 B 10.1 ± 1.0 A z One-way ANOVA analysis. Different letters within a block of a row indicate significant difference by Tukey multiple comparison method with α = Table 2. Linter, hull, and kernel summary statistics for cotton varieties produced from (as is basis) z Linter, % Hull, % Kernel, % Complete sample population ( ) Ave. ± std. dev ± ± ± 2.7 Low High ANOVA by year y ± 1.4 A 35.4 ± 2.4 B 54.1 ± 3.1 A ± 1.6 A 35.7 ± 1.1 B 54.1 ± 1.7 A ± 2.0 A 37.8 ± 1.9 A 51.2 ± 2.4 B ANOVA by location y Las Cruces 12.3 ± 2.5 A 35.8 ± 1.8 A 51.9 ± 4.0 A Lubbock 10.3 ± 1.7 B 37.0 ± 2.2 A 52.7 ± 2.7 A Stoneville 10.6 ± 1.4 AB 36.4 ± 2.1 A 53.0 ± 2.4 A z Values exclude the DP340 PIMA cultivar. y One-way ANOVA analysis. Different letters within a block of a row indicate significant difference by Tukey multiple comparison method with α = Table 3. Oil, protein, and gossypol summary statistics for cotton varieties produced from (as is basis) z Oil, % Protein, % Gossypol, % Complete sample population ( ) Ave. ± std. dev ± ± ± 0.14 Low High ANOVA by year ± 2.2 A 20.1 ± 1.4 B 0.74 ± 0.13 A ± 1.3 B 21.4 ± 2.2 A 0.65 ± 0.13 B ± 1.9 B 20.5 ± 1.0 B 0.79 ± 0.13 A ANOVA by location Las Cruces 18.6 ± 2.4 A 20.2 ± 0.9 B 0.83 ± 0.12 A Lubbock 18.1 ± 1.7 A 20.2 ± 1.3 B 0.78 ± 0.13 A Stoneville 16.7 ± 1.6 B 21.4 ± 1.8 A 0.65 ± 0.11 B z One-way ANOVA analysis. Different letters within a block of a row indicate significant difference by Tukey multiple comparison method with α = The seed-to-fiber ratio from the three years of study varied from 1.20 to 1.61 with an average value of 1.41 ± 0.11 (Table 1). This represents a marked reduction compared with values discussed in the early literature. For example, from a 1901 presentation to the Texas Cottonseed Crushers Association discussing the seed contribution to crop

4 DOWD ET AL.: SEED PROPERTIES AND COMPOSITIONAL ANALYSIS OF CURRENT COTTON CULTIVARS 63 value, McCollum uses a value of two pounds of cotton seed per pound of cotton fiber (McCollum, 1948). In his book Cottonseed Products, Thornton (1932) notes that for every 5 pounds of lint there will be produced 9 pounds of seeds, giving a seedto-fiber ratio of 1.8. Production records compiled by the Departments of Agriculture and Commerce from 1930 to 1946 yield seed-to-fiber ratios ranging between 1.60 and 1.78 (Lund, 1948). The average value over these years was The average value obtained from this study is also lower that the value of 1.5 that is often cited in current literature (e.g., Dowd, 2015) but is roughly in line with averaged values reported in recent years of the Regional High Quality (RHQ) component of the National Cotton Variety Trials (NCVT), which averaged 1.46 between 2011 and 2015 (ARS, 2017). The change in seed-to-fiber ratio appears to be due to more than simply increased fiber yield but also decreased seed mass. The average seed index of the samples within this survey was 9.8 ± 1.0 g. This result is roughly in line with recent seed indices reported as part of the RHQ-NCVT data, which averaged 9.9 g over the last five years. This compares with cultivar averages between 11 and 12 g in early studies (Garner et al., 1914; Fraps, 1916; Tharp, 1948) and an average value of 12.6 ± 1.1 from the 1964 RHQ-NCVT data, indicating that a ~10 30% reduction has occurred over the past several decades. Twelve cultivars had seed indices below 9.0 in this survey. For comparison, DP555, a cultivar that caused well known ginning difficulties, had location averaged seed indices between 7.5 and 8.7 g for the 2002 to 2007 years it was included in the RHQ-NCVT data. There was relatively little repeat or overlap data for the cultivars tested during the three years of this survey (30 cultivars in the 46 samples). However, some overlap was present. Although there was insufficient data to consider an analysis of individual cultivars by year or location, there were indications of environmental effects among the results of these samples. For instance, STV5458 grown in 2014 showed marked property differences between the Lubbock and Stoneville locations (Table S1). Given that this cultivar was developed for the Delta region, it might be expected that productivity of this cultivar would be lower in Lubbock, which was realized with a lower seed-to-fiber ratio and a dramatically smaller seed index (Table S1). Also, given the almost 30% drop in seed size and the smaller 8% reduction in seed-to-fiber ratio, it is apparent that this cultivar must have also yielded considerably less fiber in Lubbock. Although less pronounced than the differences shown for STV5458, STV4946 grown in 2015 also showed different properties in the two locations. Again the seed-fiber ratio and seed index was reduced for the Lubbock location compared with the Stoneville location (Table S2). DP1522 was produced in 2016 in both Lubbock and Las Cruces, and differences were apparent in the two samples; in this case the Lubbock sample had a greater seedto-fiber ratio and a larger seed index than did the Las Cruces sample (Table S3). Hence, significant differences in properties were observed for seed grown in different locations. In contrast with the location differences, yearto-year differences within the same cultivar were less apparent. For instance, DP340 was sampled in both 2014 and This Pima cultivar showed little difference in seed-to-fiber ratio and seed index between the two years (Tables S1 and S3). FM1944 grown in Stoneville for all three years also showed relative little variation (Tables S1 S3). STV4946 produced in Stoneville in 2015 and 2016 showed only a modest difference in seed-to-fiber ratio and no apparent difference in seed index (Tables S2 and S3). The average proportion of linters, hull, and kernel fractions appeared to be less different from early reports than were the seed index and seed-tofiber ratio results (Table 2). For instance, the average kernel mass was 52.7% compared with averages that ranged from 50.2 to 58.1% from a handful of studies spanning the period of 1906 to 1944 (Tharp, 1948). Excluding the Pima cultivar, which skewed the results due to its low levels of linters, the range of kernel values varied from 46.6 to 58.4% in this work, which is almost exactly the low and high range of values recorded from the combined early studies. The percentage of seed linters, however, appeared to be reduced compared with early reports. These reports give average values for this property ranging between 12.2 and 13.8% (Pope and Ware, 1945; Thomas and Gerdes, 1945; Martin and Thomas, 1946; Bradham and Whitten, 1948), compared with the average value of 10.7% obtained from the G. hirsutum seed. Rarely in early studies were samples noted with linter levels less than 9.0%, whereas samples with this level or less linters are quite common within this dataset. This change seems to be a result of selecting plants for increased fiber, likely promoting the retention of plants with

5 JOURNAL OF COTTON SCIENCE, Volume 22, Issue 1, 2018 improved fiber elongation genetics thereby reducing the number of non-elongated trichomes and possibly reducing the metabolic resources available to support linter development. If the percentage of linters is slightly reduced and the percentage of kernels is roughly the same, then the percentage of hull would be slightly increased. Hence, concerns of greater seed damage during ginning do not appear to be related to reduced hull mass. This, however, does not mean that the composition of hulls has not changed or that other metabolic or genetic effects have not influenced hull or seed strength. For the produced seed, the percentage of oil and protein in the kernels was typical of many prior reports. Over all of the samples, oil averaged 17.6 ± 2.0% and protein averaged 20.7 ± 1.6% on a fuzzy seed as is weight basis, compared to mean values of 17.7% oil and 20.8% protein reported by Tharp (1948). The Oil Mill Gazetteer in the 1940s reported average monthly seed oil and ammonia results from several states, and these averages typically ranged between 17 and 20% for crude oil and from 3.5 to 4.0% ammonia (or 18 21% protein), which gives a slightly greater mean value for oil and a slightly lesser value for protein than obtained in this work. Similarly in a report by Bradham and Whitten (1948), the average oil level for ten varieties grown in seven locations was 19.7% crude oil and 18.9% protein. While some early reports seem in line with current results, some reports appear to have slightly greater levels of oil and protein than found in this work. Overall, it is difficult to conclude that there has been any substantial shift in average oil or protein values due to sustained breeding efforts. Whole seed gossypol levels averaged 0.74% on the same basis, which is toward the high end of values typically measured by the authors. With an average kernel content of 53%, this translates to a kernel level of 1.4% (gossypol is frequently reported on a kernel basis), which is a little higher than the % typically observed for kernel gossypol levels. The relatively high levels appear to result because most samples included in this work were grown with irrigation, which is known to have a pronounced positive effect on seed gossypol level (Stansbury et al, 1956; Pettigrew and Dowd, 2011). The two cultivars grown under both dryland and irrigated conditions in 2015, DP1044 and NG4111, showed this effect, with irrigation increasing gossypol levels by 15 and 30%, respectively (Table S8). 64 As gossypol methods were not yet available before the late 1950s, no direct comparisons can be made for this property. DISCUSSION By comparing properties of currently grown cultivars and similar data from the early literature, it appears that the seed has changed with the continual breeding for increased fiber yield. Further evidence for this can be found in the 50 years of data from the NCVT. From any given year or even any 10 or 20 year span, variations can be large and trends difficult to discern. But when evaluated over the history of the trials, a gradual downward trend is observed for both seed-to-fiber ratio and seed index (Fig. 1). Average values from the early RHQ-NCVT data (1960s) are not far removed from the literature data discussed above, e.g., the early average seed-to-fiber ratio of around 1.7 from the trials is in the middle of the range of 1.64 and 1.78 calculated from seed and fiber production statistics of the Commerce and Agriculture Department in the 1930s and 1940s (Tharp, 1948). Averaged seed indices greater than 12 were also observed in the 1964 NCVT data, similar to early survey reports. Likewise, the data from the cultivars studied here yield ranges and average values similar to those reported in recent RHQ-NCVT results. Seed index, g Seed-to-fiber ratio Y = * X R 2 = Y = * X R 2 = Year Fig. 1. Variation in seed-to-fiber ratio and seed index from National Cotton Variety Trial data (1964 to 2015). Data is the averaged cultivar data for the Regional High Quality Trials. Standard deviations are shown on the early year data as an indication of the variation among of cultivars within the annual sampling.

6 DOWD ET AL.: SEED PROPERTIES AND COMPOSITIONAL ANALYSIS OF CURRENT COTTON CULTIVARS 65 Some care must be taken when comparing different studies and in evaluating the early literature. The literature in some cases is unclear about the basis of values (i.e., as is weight basis, dry weight basis, or at an adjusted fixed moisture level). Additionally, it can be difficult to determine if early seed indices were expressed on a whole seed or linter free basis. In some cases the basis could be figured out by context or by the magnitude of other values presented in the report, but at times this was impossible, in which case the dataset was not considered. In many cases, we relied on the data and studies noted by Tharp (1948), as it was apparent his evaluation of prior work involved similar issues, and he would have been closer to these studies and more familiar with his contemporary authors. Additionally, procedures have varied over the years. Protein, for example, was routinely determined by ammonia evolution from Kjeldahl combustion in early days, whereas combustion to nitrogen gas is standard today. Both of these approaches have the same basis, i.e., to determine the total amount of nitrogen in the sample. Hence, there should be little systemic difference due to this change. A protein-tonitrogen conversion factor of 6.25 was used in this work as it was standard for most protein estimates in the early years, although it has been argued that a conversion value of is more appropriate for cotton seed because of its relatively high level of arginine (Dowd and Wakelyn, 2010). Methods to determine crude oil have also changed, evolving from Butt-tube extractors, to repetitive immersion Soxhlet extractors or immersion-extraction systems (e.g., like a Soxtec apparatus), and even more recently to accelerated extractors under pressure. Again, the basic approach has not changed, extraction of the components soluble in hexane or petroleum ether, although there is likely more variability in these results because of different immersion times, sample preparation differences, sample moisture levels, etc. We have chosen to use a Soxtec extractor for this work, which submerges dried ground tissue under the solvent for a period of time then percolates the solvent through the sample. This approach had the advantage of being reproducible at a small scale (average reproducibility: 1 2 %), which was helpful to reduce the amount of hand dissecting of kernels needed to get good data. Making fair comparisons of seed-to-fiber ratio also required some consideration. Early estimates of seed-to-fiber ratio discussed in this work would most likely have been based on fiber and seed production numbers or experimental studies where bolls were handpicked. Also, breeding studies, which frequently report a percent fiber, e.g. as is reported in the NCVT data, are usually based on hand-picked 50 boll cotton samples. Because these samples would have been largely debris free, these numbers were used directly. In contrast, the implementation of machine harvesting in the s (Hughs, et al., 2010) dramatically increased the amounts of plant debris in field cotton samples. Because of this effect, it is difficult to estimate seed-to-fiber ratios from reported gin turn out values that range from 32 to 38%, as large amounts of trash are removed during ginning. To account for the debris in the samples used here, the ginned fiber samples were run through a Shirley cleaner. Environment and growing conditions would be expected to have some influence on the seed properties measured here, and may be reflective in the location and year differences observed in Tables 1-3. Several agronomy studies on fiber yield and properties have included results on seed index. Meredith et al (2012) report that environment accounted for 82% of the variation in seed index from seven years of NCVT data. In contrast, a number of earlier studies have noted that considerably less than half of the seed index variation (18 to 28%) was due to environment (Meredith 2003; Blanche et al., 2006, Campbell et al. 2011). In potted plants, Wang et al. (2016) noted significant smaller seed with drought conditions, while Pettigrew (2010) noted a significant difference in the seed size of irrigated plants compared with dryland plants for only two years within a four year study. Similarly, Pettigrew and Zeng (2014) noted a significant difference in seed size between irrigated and dryland plants for only one year of a four year study. Added fertilizer generally appears to increase seed index (Pettigrew and Zeng, 2014) but planting early does not (Pettigrew, 2010). Stansbury and coworkers (1956) reported on the influence of rainfall on the oil, protein, and gossypol values of the seed, noting that seed oil and gossypol levels are increased when rainfall amounts are elevated, and correspondingly protein levels are reduced. This trend seems to be present in the location data (Table 3), where the lower oil and gossypol levels and greater protein levels were observed in seed from the Stoneville location. Whether this difference was due to rainfall or irrigation or if other factors were also contributing is unclear. Finally, although the ANOVA results suggest

7 JOURNAL OF COTTON SCIENCE, Volume 22, Issue 1, 2018 differences in environment, these results are based on a limited number of varieties for the Las Cruces location, which may have influenced the results. While the average oil and protein compositional values seem little changed from prior reports, some individual compositional results were surprising. The whole seed oil level was less than 14% for two samples grown in 2016 (Table S9). DP1614, produced in Stoneville, and FM1830, produced in Lubbock, had crude oil levels of 12.7 and 13.2%, respectively, on whole seed as is basis. FM1830 was reported in the 2015 RHQ-NCVT where its oil content was 16.2%. However, from discussions with the associated analytical laboratory (Eurofins USA), we believe these numbers are reported on a linter-free basis at a reduced moisture level of about 3%, which would correspond to a 13.8% oil level on an as is whole seed basis. Hence, these values are in reasonable agreement. Low oil levels are of direct concern for oil processors but also seed producers because seed oil content is positively correlated with seedling viability (Snider et al., 2014). This trend should be monitored in the future. Smaller seed size can be particularly troublesome for ginners, who frequently take the seed as the principal part of their payment. Lower seed-to-fiber ratio translates into less seed per bale produced. For instance, at the 2017 Beltwide Cotton Conferences, multiple ginners commented that they are frequently seeing seed yields less than 600 pounds per bale, which roughly translates to seed-to-fiber recovery ratio of 1.2 (assuming a 500 pound bale) down from amounts closer to 800 pounds per bale or a seed-tofiber recovery ratio of 1.6. Ginners believe that additional seed loss might be occurring as they have observed increased seed exiting the gin in the motes and fiber cleaning operations. Because the 1.2 seed-to-fiber recovery ratio is lower than the current average seed-to-fiber ratio of 1.4, additional factors may be affecting seed yields. Small seed may be capable of passing through the ribs of gin stands, and would then be lost during subsequent lint cleaning operations. If this is occurring, then it seems that there may be a critical seed size that results in accelerated losses. At present, only a rough estimate can be made of this size, but given the problems with the problematic DP555 cultivar with average RHQ-NCVT seed indices between 7.5 and 8.7 g, the number of current cultivars with seed indices <9 g, and the current complaints of ginners, this critical seed index size may be around 8.5 g. 66 Damaged or broken seed may also contribute to losses. Greater seed damage may result from inherently weaker seed, and cultivars prone to mechanical damage are known. However, if weaker seed are produced today, it is not apparent from the percentage of seed hull, which is roughly the same as in early reports. Increased gin speed is well known to increase seed damage (Delouche, 1986). With the pressure to gin seed cotton faster, ginning speed may also be a contributing factor to increased seed damage and accelerated seed loss. SUMMARY Seed cotton varieties from the growing seasons were evaluated for a number of seed-fiber traits and compositional properties. Seed-fiber ratio averaged 1.41 ± 0.11, which is dramatically lower than values reported from early reports. Correspondingly, seed index has also fallen, indicating that the breeding of cotton for increased fiber yield appears to have resulted in substantive changes in seed size. A reduction in the amount of cottonseed linters was also apparent. Significant changes in the proportions of the hull and kernel fractions, and in the seed oil and protein levels were not observed. The scale of the changes to the seed-tofiber ratio and seed index appears to have significant detrimental effects for ginning operations. ACKNOWLEDGEMENTS The authors thank Rick Byler (ARS), Greg Holt (ARS) and Ed Hughs (ARS) for providing seed cotton samples and Ellen Keene (ARS) for help in collecting early NCVT data. REFERENCES ARS, Standard procedures for foreign matter and moisture analytical tests used in cotton ginning research. Agricultural Handbook No. 422, Agricultural Research Service, U.S. Department of Agriculture. 13 pp. ARS, stoneville-ms/crop-genetics-research/docs/nationalcotton-variety-test/ AOCS, Official methods and recommended practices of the American Oil Chemists Society. (Firestone, D., ed.) 5 th edition, AOCS Press, Champaign, IL.

8 DOWD ET AL.: SEED PROPERTIES AND COMPOSITIONAL ANALYSIS OF CURRENT COTTON CULTIVARS 67 Blanche, S.B., Myers, G.O., Zumba, J.Z., Caldwell, D.W. and Hayes, J Stability comparisons between conventional and near-isogenic transgenic cotton cultivars. J. Cotton Sci. 10: Bradham, D.A., and Whitten, M.E Factors affecting milling quality of cottonseed and means for improving quality. Oil Mill Gazet. 52(8): Campbell, B.T., Chee, P.W., Lubbers, E., Bowman, D.T., Meredith, Jr., W.R., Johnson, J.T., Fraser, D., Bridges, W., and Jones, D.C Dissecting genotype x environment interactions and trait correlations in the Pee Dee Cotton Germplasm Collection following seventy years of plant breeding. Crop Sci. 51: Delouche, J.C Harvest and post-harvest factors affecting the quality of cotton planting seed and seed quality evaluation. p In J.R. Mauney, and Steward, J. McD. (eds.) Cotton Physiology. The Cotton Foundation Reference Book Series, No. 1, The Cotton Foundation, Memphis, TN. Dowd, M.K Seed. p In Fang, D.D., and Percy, R.G. (eds.) Cotton. Agronomy Series No. 57, American Society of Agronomy, Crop Science of America, and Soil Science of America, Madison, WI. Dowd, M.K., and Wakelyn, P.J Cottonseed: current and future utilization. p In Wakelyn, P.J., and Chaudhry, M.R. (eds.) Cotton: Technology for the 21st Century. International Cotton Advisory Committee, Washington, DC. Fraps, G.S The composition of cotton seed meal and cotton seed. Texas Agricultural Experiment Station, Bulletin #189, Garner, W.W., Allard, H.A., and Foubert, C.L Oil content of seeds as affected by the nutrition of the plant. J. Agric. Res. 3, Hughs, S.E., Parnell, Jr., C.B., and Wakelyn, P.J Cotton harvesting and ginning in the 21 st century. p In Wakelyn, P.J., and Chaudhry, M.R. (eds.) Cotton: Technology for the 21st Century. International Cotton Advisory Committee, Washington, DC. Lund, C.E., Production and consumption of cottonseed and cottonseed products. p In Bailey, A.E. (ed.) Cottonseed and Cottonseed Products. Interscience, NY. Martin, W.J., and Thomas, H.L Progress in cotton seed quality. Oil Mill Gazet. 51(1): McCollum, A.R What the cottonseed oil mill has done for the farmer. Oil Mill Gazet. 52(11):27, Meredith, Jr., W.R Thirty-six years of Regional High Quality Variety Tests. p In Proc. Beltwide Cotton Conf., Nashville, TN, 6-10 Jan National Cotton Council of America, Memphis, TN. Meredith, Jr., W.R., Boykin, D.L., Bourland, F.M., Caldwell, W.D., Campbell, B.T., Gannaway, J.R., Glass, K., Jones, A.P., May, L.M., Smith, C.W., and Zhang, J Genotype x environment interactions over seven years for yield, yield components, fiber quality, and gossypol traits on the Regional High Quality Tests. J. Cotton Sci. 16: Pettigrew, W.T Impact of varying planting dates and irrigation regimes on cotton growth and lint yield production. Agron. J. 102(5): Pettigrew, W.T., and Dowd, M.K Varying planting dates or irrigation regimes alters cottonseed composition. J. Cotton Sci. 15: Pettigrew, W.T. and Zeng, L Interactions among irrigation and nitrogen fertilization regimes on mid-south cotton production. Agron. J. 106(5): Pope, O.A., and Ware, J.O Effect of variety, location, and season on oil, protein, and fuzz of cottonseed and on fiber properties of lint. Technical Bulletin No. 903, U.S. Dept. of Agriculture Snider, J.L., Collins, G.D., Whitaker, J., Chapman, K.D., Horn, P., and Grey, T.L Seed size and oil content are key determinants of seedling vigor in Gossypium hirsutum. J. Cotton Sci. 18:1-9. Stansbury, M.F., Pons, Jr., W.A., and Den Hartog, G.T Relations between oil, nitrogen, and gossypol in cottonseed kernels. J. Amer. Oil Chem. Soc. 33: Tharp, W.H Cottonseed composition--relation to variety, maturity and environment of the plant. p In Bailey, A.E. (ed.) Cottonseed and Cottonseed Products, Interscience, NY. Thomas, H.L., and Gerdes, F.L Cotton seed quality test results of interest to cottonseed oil mill superintendents. Oil Mill Gazet. 50(3):23, Thornton, M.K Cottonseed Products, Oil Mill Gazetteer, Wharton, TX. USDA, Agricultural Marketing Service Cotton and Tobacco Program, Cotton Varieties Planted 2014 Crop Report, Memphis, TN, 10 pp. USDA, Agricultural Marketing Service Cotton and Tobacco Program, Cotton Varieties Planted 2015 Crop Report, Memphis, TN, 10 pp. USDA, Agricultural Marketing Service Cotton and Tobacco Program, Cotton Varieties Planted 2016 Crop Report, Memphis, TN, 10 pp. Wang, R., Ji, S., Zhang, P., Meng, Y., Wang, Y., Chen, B., and Zhou, Z Drought effects on cotton yield and fiber quality on different fruiting branches. Crop Sci. 56:

9 JOURNAL OF COTTON SCIENCE, Volume 22, Issue 1, Table S1. Seed/fiber ratio and seed indices for cotton cultivars grown in 2014 (as is basis). Variety - Location Dryland / Irrigated Seed/fiber ratio Seed index, g DP340 - PIMA - Las Cruces Irrigated 1.56 ± ± 0.03 DP Stoneville Irrigated 1.54 ± ± 0.25 DP Lubbock Irrigated 1.37 ± ± 0.11 DP Stoneville Irrigated 1.28 ± ± 0.33 FM Stoneville Irrigated 1.54 ± ± 0.25 FM Lubbock Irrigated 1.23 ± ± 0.13 PHY339 - Las Cruces Irrigated 1.57 ± ± 0.22 PHY565 - Las Cruces Irrigated 1.50 ± ± 0.08 STV Stoneville Irrigated 1.39 ± ± 0.33 STV Lubbock Irrigated 1.40 ± ± 0.06 STV Stoneville Irrigated 1.51 ± ± 0.11 Table S2. Seed/fiber ratio and seed indices for cotton cultivars grown in 2015 (as is basis). Variety - Location Dryland / Irrigated Seed/fiber ratio Seed index, g Acala Las Cruces Irrigated 1.60 ± ± 0.15 DP Lubbock Dryland 1.32 ± ± 0.08 DP Lubbock Irrigated 1.47 ± ± 0.09 DP Stoneville Irrigated 1.39 ± ± 0.10 FM Stoneville Irrigated 1.56 ± ± 0.08 FM Lubbock Dryland 1.44 ± ± 0.14 NG Lubbock Dryland 1.33 ± ± 0.09 NG Lubbock Irrigated 1.46 ± ± 0.03 PHY444 - Stoneville Irrigated 1.31 ± ± 0.18 PHY499 - Stoneville Irrigated 1.34 ± ± 0.12 STV Lubbock Irrigated 1.46 ± ± 0.15 STV Stoneville Irrigated 1.52 ± ± 0.10 STV Stoneville Irrigated 1.54 ± ± 0.06 Table S3. Seed/fiber ratio and seed indices for cotton cultivars grown in 2016 (as is basis). Variety - Location Dryland / Irrigated Seed/fiber ratio Seed index, g CP Lubbock Irrigated 1.38 ± ± 0.07 DG3385 Las Cruces Irrigated 1.31 ± ± 0.24 DP340 - PIMA - Las Cruces Irrigated 1.61 ± ± 0.13 DP Stoneville Irrigated 1.37 ± ± 0.08 DP Las Cruces Irrigated 1.26 ± ± 0.07 DP1522 Lubbock Irrigated 1.44 ± ± 0.11 DP Lubbock Irrigated 1.41 ± ± 0.07 DP Stoneville Irrigated 1.20 ± ± 0.07 FM Lubbock Irrigated 1.28 ± ± 0.05 FM Lubbock Irrigated 1.26 ± ± 0.23 FM Stoneville Irrigated 1.51 ± ± 0.14 NG Lubbock Irrigated 1.33 ± ± 0.09 NG Lubbock Irrigated 1.34 ± ± 0.01 NG Lubbock Irrigated 1.43 ± ± 0.12 NG Lubbock Irrigated 1.36 ± ± 0.10 PHY243 - Lubbock Irrigated 1.45 ± ± 0.05 PHY308 - Lubbock Irrigated 1.49 ± ± 0.14 PHY333 - Lubbock Irrigated 1.39 ± ± 0.13

10 DOWD ET AL.: SEED PROPERTIES AND COMPOSITIONAL ANALYSIS OF CURRENT COTTON CULTIVARS 69 Table S3. (continued) Variety - Location Dryland / Irrigated Seed/fiber ratio Seed index, g PHY444 - Stoneville Irrigated 1.28 ± ± 0.17 PHY499 - Stoneville Irrigated 1.30 ± ± 0.21 STV Stoneville Irrigated 1.43 ± ± 0.16 STV Stoneville Irrigated 1.46 ± ± 0.02 Table S4. Proportion of cottonseed linter, hull and kernel tissues from cotton cultivars grown in 2014 (as is basis). Variety - Location Dryland / Irrigated Linter Hull Kernel DP340 - PIMA - Las Cruces Irrigated 3.1 ± ± ± 0.5 DP Stoneville Irrigated 10.8 ± ± ± 0.5 DP Lubbock Irrigated 12.6 ± ± ± 0.1 DP Stoneville Irrigated 10.7 ± ± ± 0.2 FM Stoneville Irrigated 9.7 ± ± ± 0.3 FM Lubbock Irrigated 9.7 ± ± ± 0.6 PHY339 - Las Cruces Irrigated 9.3 ± ± ± 0.3 PHY565 - Las Cruces Irrigated 13.3 ± ± ± 0.5 STV Stoneville Irrigated 10.1 ± ± ± 0.6 STV Lubbock Irrigated 8.8 ± ± ± 0.4 STV Stoneville Irrigated 10.5 ± ± ± 0.3 Table S5. Proportion of cottonseed linter, hull, kernel tissues from cotton cultivar grown in 2015 (as is basis). Variety - Location Dryland / Irrigated Linter Hull Kernel Acala Las Cruces Irrigated 9.9 ± ± ± 0.4 DP Lubbock Dryland 11.5 ± ± ± 0.4 DP Lubbock Irrigated 12.6 ± ± ± 0.5 DP Stoneville Irrigated 11.4 ± ± ± 0.5 FM Stoneville Irrigated 11.2 ± ± ± 0.6 FM Lubbock Dryland 10.4 ± ± ± 0.4 NG Lubbock Dryland 8.6 ± ± ± 0.4 NG Lubbock Irrigated 9.7 ± ± ± 0.3 PHY444 - Stoneville Irrigated 6.4 ± ± ± 0.4 PHY499 - Stoneville Irrigated 9.3 ± ± ± 0.3 STV Lubbock Irrigated 11.0 ± ± ± 0.2 STV Stoneville Irrigated 10.4 ± ± ± 0.2 STV Stoneville Irrigated 9.7 ± ± ± 0.2 Table S6. Proportion of cottonseed linter, hull, kernel tissues from cotton cultivars grown in 2016 (as is basis). Variety - Location Dryland / Irrigated Linter Hull Kernel CP Lubbock Irrigated 9.8 ± ± ± 0.5 DG3385 Las Cruces Irrigated 14.6 ± ± ± 0.7 DP340 - PIMA - Las Cruces Irrigated 3.8 ± ± ± 0.1 DP Stoneville Irrigated 12.9 ± ± ± 0.3 DP1522 Las Cruces Irrigated 14.4 ± ± ± 0.6 DP Lubbock Irrigated 11.4 ± ± ± 1.0 DP Lubbock Irrigated 10.0 ± ± ± 0.6 DP Stoneville Irrigated 11.1 ± ± ± 0.4 FM Lubbock Irrigated 13.3 ± ± ± 0.2 FM Lubbock Irrigated 6.2 ± ± ± 0.6 FM Stoneville Irrigated 11.9 ± ± ± 0.6 NG Lubbock Irrigated 11.4 ± ± ± 0.6

11 JOURNAL OF COTTON SCIENCE, Volume 22, Issue 1, Table S6. (continued) Variety - Location Dryland / Irrigated Linter Hull Kernel NG Lubbock Irrigated 11.1 ± ± ± 0.3 NG Lubbock Irrigated 9.5 ± ± ± 0.2 NG Lubbock Irrigated 9.5 ± ± ± 0.3 PHY243 - Lubbock Irrigated 8.4 ± ± ± 0.6 PHY308 - Lubbock Irrigated 11.6 ± ± ± 0.3 PHY333 - Lubbock Irrigated 8.6 ± ± ± 0.6 PHY444 - Stoneville Irrigated 10.6 ± ± ± 0.2 PHY499 - Stoneville Irrigated 12.0 ± ± ± 0.4 STV Stoneville Irrigated 11.2 ± ± ± 0.6 STV Stoneville Irrigated 11.5 ± ± ± 0.1 Table S7. Percentage of seed oil, protein, and gossypol for 2014 cultivars (as is basis). Variety - Location Dryland / Irrigated Oil Protein Gossypol DP340 - PIMA - Las Cruces Irrigated 22.1 ± ± ± 0.01 DP Stoneville Irrigated 18.2 ± ± ± 0.02 DP Lubbock Irrigated 17.4 ± ± ± 0.01 DP Stoneville Irrigated 16.9 ± ± ± 0.02 FM Stoneville Irrigated 15.8 ± ± ± 0.00 FM Lubbock Irrigated 22.4 ± ± ± 0.01 PHY339 - Las Cruces Irrigated 20.4 ± ± ± 0.01 PHY565 - Las Cruces Irrigated 16.2 ± ± ± 0.01 STV Stoneville Irrigated 17.4 ± ± ± 0.02 STV Lubbock Irrigated 20.4 ± ± ± 0.02 STV Stoneville Irrigated 19.3 ± ± ± 0.02 Table S8. Percentage of seed oil, protein, and gossypol for 2015 cultivars (as is basis). Variety - Location Dryland / Irrigated Oil Protein Gossypol Acala Las Cruces Irrigated 18.3 ± ± ± 0.04 DP Lubbock Dryland 18.6 ± ± ± 0.02 DP Lubbock Irrigated 17.7 ± ± ± 0.02 DP Stoneville Irrigated 16.1 ± ± ± 0.03 FM Stoneville Irrigated 15.0 ± ± ± 0.01 FM Lubbock Dryland 18.4 ± ± ± 0.01 NG Lubbock Dryland 18.2 ± ± ± 0.02 NG Lubbock Irrigated 18.2 ± ± ± 0.00 PHY444 - Stoneville Irrigated 17.3 ± ± ± 0.02 PHY499 - Stoneville Irrigated 16.7 ± ± ± 0.02 STV Lubbock Irrigated 18.1 ± ± ± 0.03 STV Stoneville Irrigated 19.6 ± ± ± 0.02 STV Stoneville Irrigated 15.7 ± ± ± 0.02 Table S9. Percentage of seed oil, protein, and gossypol for 2016 cultivars (as is basis). Variety - Location Dryland / Irrigated Oil Protein Gossypol CP Lubbock Irrigated 18.2 ± ± ± 0.00 DG3385 Las Cruces Irrigated 17.0 ± ± ± 0.04 DP340 - PIMA - Las Cruces Irrigated 21.0 ± ± ± 0.00 DP Stoneville Irrigated 16.3 ± ± ± 0.00 DP1522 Las Cruces Irrigated 15.5 ± ± ± 0.03 DP Lubbock Irrigated 16.2 ± ± ± 0.02

12 DOWD ET AL.: SEED PROPERTIES AND COMPOSITIONAL ANALYSIS OF CURRENT COTTON CULTIVARS 71 Table S9. (continued) Variety - Location Dryland / Irrigated Oil Protein Gossypol DP Lubbock Irrigated 17.4 ± ± ± 0.02 DP Stoneville Irrigated 12.7 ± ± ± 0.01 FM Lubbock Irrigated 13.2 ± ± ± 0.01 FM Lubbock Irrigated 18.8 ± ± ± 0.02 FM Stoneville Irrigated 15.4 ± ± ± 0.01 NG Lubbock Irrigated 16.6 ± ± ± 0.02 NG Lubbock Irrigated 16.4 ± ± ± 0.01 NG Lubbock Irrigated 19.0 ± ± ± 0.00 NG Lubbock Irrigated 18.1 ± ± ± 0.01 PHY243 - Lubbock Irrigated 19.8 ± ± ± 0.02 PHY308 - Lubbock Irrigated 17.7 ± ± ± 0.02 PHY333 - Lubbock Irrigated 18.8 ± ± ± 0.02 PHY444 - Stoneville Irrigated 16.7 ± ± ± 0.02 PHY499 - Stoneville Irrigated 16.7 ± ± ± 0.02 STV Stoneville Irrigated 17.9 ± ± ± 0.02 STV Stoneville Irrigated 17.0 ± ± ± 0.02 Table S10. Seed/fiber ratio and seed indices for cotton cultivars grown in 2014 (dry weight basis). Variety - Location Dryland / Irrigated Seed/fiber ratio Seed index, g DP340 - PIMA - Las Cruces Irrigated 1.54 ± ± 0.01 DP Stoneville Irrigated 1.50 ± ± 0.20 DP Lubbock Irrigated 1.33 ± ± 0.10 DP Stoneville Irrigated 1.26 ± ± 0.29 FM Stoneville Irrigated 1.50 ± ± 0.20 FM Lubbock Irrigated 1.21 ± ± 0.11 PHY339 - Las Cruces Irrigated 1.58 ± ±0.21 PHY565 - Las Cruces Irrigated 1.49 ± ± 0.05 STV Stoneville Irrigated 1.36 ± ± 0.29 STV Lubbock Irrigated 1.37 ± ± 0.05 STV Stoneville Irrigated 1.48 ± ± 0.05 Table S11. Seed/fiber ratio and seed indices for cotton cultivars grown in 2015 (dry weight basis). Variety - Location Dryland / Irrigated Seed/fiber ratio Seed index, g Acala Las Cruces Irrigated 1.57 ± ± 0.13 DP Lubbock Dryland 1.28 ± ± 0.07 DP Lubbock Irrigated 1.42 ± ± 0.08 DP Stoneville Irrigated 1.33 ± ± 0.08 FM Stoneville Irrigated 1.50 ± ± 0.08 FM Lubbock Dryland 1.39 ± ± 0.12 NG Lubbock Dryland 1.29 ± ± 0.07 NG Lubbock Irrigated 1.43 ± ± 0.02 PHY444 - Stoneville Irrigated 1.26 ± ± 0.15 PHY499 - Stoneville Irrigated 1.28 ± ± 0.11 STV4946 Lubbock Irrigated 1.42 ± ± 0.14 STV Stoneville Irrigated 1.47 ± ± 0.10 STV Stoneville Irrigated 1.48 ± ± 0.05

13 JOURNAL OF COTTON SCIENCE, Volume 22, Issue 1, Table S12. Seed/fiber ratio and seed indices for cotton cultivars grown in 2016 (dry weight basis). Variety - Location Dryland / Irrigated Seed/fiber ratio Seed index, g CP3475 Lubbock Irrigated 1.35 ± ± 0.06 DG3385 Las Cruces Irrigated 1.28 ± ± 0.21 DP340 - PIMA - Las Cruces Irrigated 1.57 ± ± 0.12 DP Stoneville Irrigated 1.34 ± ± 0.07 DP Las Cruces Irrigated 1.22 ± ± 0.06 DP1522 Lubbock Irrigated 1.40 ± ± 0.10 DP1612 Lubbock Irrigated 1.38 ± ± 0.07 DP Stoneville Irrigated 1.16 ± ± 0.06 FM1830 Lubbock Irrigated 1.24 ± ± 0.05 FM1911 Lubbock Irrigated 1.23 ± ± 0.21 FM Stoneville Irrigated 1.47 ± ± 0.12 NG3405 Lubbock Irrigated 1.30 ± ± 0.09 NG3406 Lubbock Irrigated 1.31 ± ± 0.01 NG3517 Lubbock Irrigated 1.39 ± ± 0.12 NG4545 Lubbock Irrigated 1.33 ± ± 0.09 PHY243 Lubbock Irrigated 1.43 ± ± 0.04 PHY308 Lubbock Irrigated 1.46 ± ± 0.13 PHY333 Lubbock Irrigated 1.37 ± ± 0.12 PHY444 Stoneville Irrigated 1.25 ± ± 0.15 PHY499 Stoneville Irrigated 1.27 ± ± 0.19 STV Stoneville Irrigated 1.40 ± ± 0.15 STV Stoneville Irrigated 1.42 ± ± 0.01 Table S13. Proportion of cottonseed linter, hull and kernel tissues from cotton varieties grown in 2014 (dry weight basis). Variety Linter Hull Kernel DP340 PIMA - Las Cruces Irrigated 3.1 ± ± ± 0.7 DP Stoneville Irrigated 11.0 ± ± ± 0.6 DP Lubbock Irrigated 12.7 ± ± ± 0.2 DP Stoneville Irrigated 10.9 ± ± ± 0.1 FM Stoneville Irrigated 9.9 ± ± ± 0.5 FM Lubbock Irrigated 9.8 ± ± ± 0.4 PHY339 - Las Cruces Irrigated 9.4 ± ± ± 0.3 PHY565 - Las Cruces Irrigated 13.5 ± ± ± 0.3 STV Stoneville Irrigated 10.3 ± ± ± 0.7 STV Lubbock Irrigated 8.9 ± ± ± 04 STV Stoneville Irrigated 10.6 ± ± ± 0.2 Table S14. Proportion of cottonseed linter, hull, kernel tissues from cotton varieties grown in 2015 (dry weight basis). Variety Linter Hull Kernel Acala Las Cruces Irrigated 10.0 ± ± ± 0.5 DP Lubbock Dryland 11.6 ± ± ± 0.36 DP Lubbock Irrigated 12.6 ± ± ± 0.5 DP Stoneville Irrigated 11.5 ± ± ± 0.5 FM Stoneville Irrigated 11.3 ± ± ± 0.6 FM Lubbock Dryland 10.5 ± ± ± 0.4 NG Lubbock Dryland 8.6 ± ± ± 0.5 NG Lubbock Irrigated 9.7 ± ± ± 0.3 PHY444 - Stoneville Irrigated 6.5 ± ± ± 0.4 PHY499 - Stoneville Irrigated 9.4 ± ± ± 0.3